Release Strategy Proposal

[Martin Graesslin]

On Friday 26 April 2013 18:36:07 Scott Kitterman wrote:
> On Saturday, April 27, 2013 12:27:49 AM Aaron J. Seigo wrote:
> > On Friday, April 26, 2013 18:15:03 Scott Kitterman wrote:
> > > impression is that your goal is to get more people working on KFS/PW2
> > > and
> > > your chosen method to achieve it is by enforcing a stop on feature work
> > > in
> > > KDE4 platform/workspace.
> > 
> > this isn't our goal. our goal is to allow those of us currently working on
> > workspaces to focus feature development where it already is: the Qt5 / KF5
> > ports of Plasma (you can see the work in the plasma-framework repository)
> > and kde-workspace.
> > 
> > we simply can't do both 4.x and 5.x feature development, for both
> > technical
> > and man-power reasons. we also don't want to let the 4.x releases of kde-
> > workspace happen without any visible changes (have fun writing those
> > release notes :) without some communication behind it, and we also
> > understand there is a desire from at least some of our downstreams to
> > have a long-term release they can focus on. so this represents a
> > potential win-win situation in which we can communicate our shift of
> > focus while also delivering a long-term release.
> > 
> > at the same time, we don't want to put pressure on application developers
> > to shift focus away from 4.x. until KF5 is ready, there is no reason to
> > do this.
> > 
> > we feel that stating clearly that workspaces is feature frozen covers all
> > the above concerns.
> > 
> > if this also causes other people to join our efforts .. that would be a
> > bonus. it is not the primary goal, however.
> > 
> > > Avoiding a long period of feature freeze between with KDE4
> > > platform/workspace development stops and when KFS/PW2 is usable and
> > > available is an important part of the story.
> > 
> > does it matter if the applications keep releasing? has anyone noticed this
> > is already the case for a year in kdelibs?
> 
> It matters less that workspace is frozen if applications aren't, but unlike
> kdelibs, users see the workspace.  I just upgraded my main laptop to 4.10.2
> and I appreciate the changes in both (4.10 continues the trend of every
> release starting with 4.1 being a significant improvement over the rest).
> 
> I'm not opposed to the suggestion, only trying to suggest another way to
> look at things.
Let's face it: I won't be able to port KWin to Qt 5 if we continue to do 
releases based on Qt 4. I tried it: the changes are too large to have it in 
one source base with compile switches. So either master switches to Qt 5 or we 
will stick with Qt 4. And it's not like we run a script on KWin source base 
and are done. We need at least half a year of a safety net to find all the 
regressions we will introduce.

The situation right now in the workspaces is that Plasma is already kind of 
frozen as feature development moved to plasma-frameworks. Improvements are 
only available in KWin due to the Qt 5 porting. As soon as 4.11 is feature 
frozen the work has to shift to Qt 5 for KWin as well, otherwise we will not 
able to do the transition. So even if we did a 4.12 for workspaces it would 
not include any new features in the most visible areas.

Hmm seems like I just explained the technical and manpower reasons Aaron 
mentioned :-)

Cheers
Martin